Combination of an ACLY inhibitor with a GLP-1R agonist exerts additive benefits on nonalcoholic steatohepatitis and hepatic fibrosis in mice

Abstract

Increased liver de novo lipogenesis (DNL) is a hallmark of nonalcoholic steatohepatitis (NASH). A key enzyme controlling DNL upregulated in NASH is ATP citrate lyase (ACLY). In mice, inhibition of ACLY reduces liver steatosis, ballooning, and fibrosis and inhibits activation of hepatic stellate cells. Glucagon-like peptide-1 receptor (GLP-1R) agonists lower body mass, insulin resistance, and steatosis without improving fibrosis. Here, we find that combining an inhibitor of liver ACLY, bempedoic acid, and the GLP-1R agonist liraglutide reduces liver steatosis, hepatocellular ballooning, and hepatic fibrosis in a mouse model of NASH. Liver RNA analyses revealed additive downregulation of pathways that are predictive of NASH resolution, reductions in the expression of prognostically significant genes compared with clinical NASH samples, and a predicted gene signature profile that supports fibrosis resolution. These findings support further investigation of this combinatorial therapy to treat obesity, insulin resistance, hypercholesterolemia, steatohepatitis, and fibrosis in people with NASH.

Keywords: ACLY; GLP-1R agonist; MASH; NAFLD; NASH; bempedoic acid; combination treatment; fatty acid metabolism; hepatic fibrosis; lipogenesis; liraglutide; mouse model; semaglutide.

Graphical abstract

Figure 1

Lira and BemA lower body mass, adiposity, insulin sensitivity, and serum cholesterol without increasing serum triglycerides (A and B) Percentage change in body weight (A) and change in adiposity (post-pre) (B) throughout intervention. (C–E) Intraperitoneal glucose tolerance test (GTT) (1.25 g/kg) (C) at 4 week intervention, intraperitoneal (i.p.) insulin tolerance test (ITT) (1.3 U/kg) (D) at 4 week intervention, and i.p. pyruvate tolerance test (PTT) (1.5 g/kg) (E) at 5 week intervention with time plots and area under the curve (AUC). (F) Fasted serum insulin collected via tail nick near end of intervention (9 weeks). (G and H) Fed serum cholesterol (G) from blood collected by cardiac puncture at sacrifice and fasted serum triglycerides (H). Data are means ± SEM. Colored bars signify comparisons between groups and vehicle. Significance was accepted at p <0.05 and determined via one-way ANOVA or repeated-measures two-way ANOVA with Tukey post hoc, where appropriate. White circles are individual mice per group (n = 8–9 mice/group). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. Vehicle (saline treatments subcutaneously every 2 days), Lira (70 μg/kg Lira subcutaneously every 2 days before lights out), and Lira+BemA (BemA 10 mg/kg in diet and 70 μg/kg Lira subcutaneously every 2 days before lights out).

Figure 2

Lira and BemA reduce liver steatosis, ballooning, inflammation, and fibrosis (A) Liver fat percentage as measured by time-domain nuclear magnetic resonance (NMR). (B) Liver triglycerides. (C–H) Representative micrographs of H&E- (top) and picrosirius red- (PSR; bottom) stained sections (10×) (C) along with histograms of histological grades of liver steatosis (D), hepatocellular ballooning (E), lobular inflammation (F), and composite NAFLD activity score (NAS) (G). (H and I) Percentage of positive PSR area (H) and parts of whole indicating presence of moderate, zone 3 perisinusoidal fibrosis (I). Data are means ± SEM. Colored bars signify comparisons between groups and vehicle with percentages listed next to them. Significance was accepted at p <0.05 and determined via one-way ANOVA with Tukey post hoc, or, for histological score analysis, a Kruskal-Wallis test was used with Dunn’s post hoc test to correct for multiple comparisons, where appropriate. White circles are individual mice per group (n = 8–9 mice/group). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. Vehicle (saline treatments subcutaneously every 2 days), Lira (70 μg/kg Lira subcutaneously every 2 days before lights out), and Lira+BemA (BemA 10 mg/kg in diet and 70 μg/kg Lira subcutaneously every 2 days before lights out).

Figure 3

Gene set analysis reveals additive downregulation of fibrosis-related pathways by combination treatment that are associated with NASH resolution (A) Signature scores of transcriptional pathways most affected by combination treatment. Each heatmap column represents an individual sample, along with rows annotated according to treatment cohort, histology, and gene set scores. (B) PCA of control, monotherapy, and combination treatments based on gene set scores. (C) Odds ratio and 95% confidence interval associated with hepatic steatosis, ballooning degeneration, inflammation, fibrosis, NAS, and liver adiposity measurements based on PC1 of gene set scores. (D and E) Gene expression of hepatic stellate cell markers (D) and chemokines (E) associated with NASH progression. Black circles are individual mice per group (n = 5–9 mice/group). Black circles represent individual mice per group (n = 5–9 mice/group). Boxplots show median and interquartile range, trailing lines represent 95% confidence interval. Difference between groups were assessed by one-way ANOVA followed by Dunnett’s post hoc test using the control group as the reference level. Significance was accepted at p < 0.0033 to correct for Bonferroni multiple hypothesis testing.

Figure 4

Combination treatment promotes gene expression associated with NASH resolution (A) Hierarchical clustering of healthy human patients, patients with NAFLD/NASH, and experimental cohorts based on the standardized expression of prognostically significant orthologous genes involved in NASH progression. (B) Classification of human-derived samples based on similarity to combination treatment gene signature identifies differential enrichment of parenchymal and nonparenchymal cell types between predicted classes. Boxplots show median and interquartile range, trailing lines represent 95% confidence interval. Outliers are represented by black circles. Difference between groups were assessed by Students’ t test and significance was accepted at p < 0.0056 to correct for Bonferroni multiple hypothesis testing.